Research on GPU Cluster acceleration of FDTD algorithm

Author

Zhang, Bo ; Xue, Zhenghui ; Ren, Wu ; Li, Weiming ; Sheng, Xinqing

Author_Institution

Sch. of Inf. & Electron., Beijing Inst. of Technol., Beijing, China

Volume

1

fYear

2012

fDate

5-8 May 2012

Firstpage

1

Lastpage

4

Abstract

In this paper, the Graphics Processing Unit(GPU) computing together with Cluster of Workstations(COW) techniques are discussed. Based on which a high efficiency hardware accelerated Finite-Difference Time-Domain (FDTD) algorithm has been developed and implemented in two nodes COW system. Two typical EM simulation examples have been simulated. The results show good agreement between the GPU+COW scheme and the traditional CPU computing while at least 60 times speedup has been achieved.

Keywords

finite difference time-domain analysis; graphics processing units; mathematics computing; workstation clusters; CPU computing; EM simulation; FDTD algorithm; GPU cluster acceleration; finite-difference time-domain algorithm; graphics processing unit; workstation cluster; Acceleration; Finite difference methods; Graphics processing unit; Instruction sets; Performance evaluation; Resource management; Time domain analysis; COW; FDTD; GPU; Hardware Accelerating;

fLanguage

English

Publisher

ieee

Conference_Titel

Microwave and Millimeter Wave Technology (ICMMT), 2012 International Conference on

Conference_Location

Shenzhen

Print_ISBN

978-1-4673-2184-6

Type

conf

DOI

10.1109/ICMMT.2012.6229926

Filename

6229926